 Hi, everybody. I would like to talk to you about the states of the comments, meaning the comment feature or methodologies within Next Cloud. So in version 9.0 back then, we introduced this feature that you could comment on files. And we designed it carefully to be modular and to be reusable. So basically, that was split in three parts. For one, the public API that anybody else could implement as well. For the other, our own implementation using the database and the GUI that is implemented as an app that really would show the sidebar in the files app. To the API somewhat also belongs a functionality so that the comments are hooked into the DAF functionality. So in fact, you can read them from web DAF and write them from web DAF. So with this in mind, it was well integrated into the different parts of Next Cloud. And like you see here now, these are against examples of the DAF connection, web DAF connection, and the separation into the GUI. So with that in mind, we also wanted first to enable maybe other implementers who have a fancy file system to be able to write the comments directly into the file and not into the database like our implementation does while still keeping the possibility of functionality to work via web DAF. And the web DAF way is actually also the way how the GUI works when talking to the server. And the other idea was also to be able to use a comments infrastructure in other applications or use cases. For instance, you could do a like feature that would just use the comments infrastructure. Or we have the signature panel and planning that would also make use of this. However, how is the adoption of the comments? Is anybody here in the room really using it? Hands up? OK, there are a few. That's a sad. Do you use it also for really discussing with other users or more like just for yourself? So who's discussing? Who's discussing about files? OK, OK. Good, that's not that many. I think that's the issue. Because I don't see very much of public talking about that feature itself. And my own dog footing or what I saw in our corporate instance is that it's kind of not really used. And the main problem is that it's very hard kind of to pull other users in. Because it's, for instance, integrated with the activity. So whether new comments are written, it appears on the activity stream. But there are no notifications. So you cannot mention a user. And that's why, yes, my use case here is I also use it. But mostly kind of just to say, yeah, I listen to that podcast until I listen to that point so I can continue there, which is not as much as you could do. So to improve the whole thing, yes, you mentioned you want to do this. And in fact, I think almost half a year ago or so, I already started to implement this. But on the way things happened and now it's the time to work on this. So we have a four step approach on how to realize this. Because the comments themselves, they just take plain text and they don't do anything else. So when we want to do user mentioning, we want kind of to put their username somehow. And unlike GitHub, for instance, we have the separation between a user ID, which can be very, very crude. If it's the default LAP setting, for instance, and we have the display name. So it kind of is a proper formatted readable name. And that should appear there. So yeah, that's why it's a four step approach. So first we want to do a very simple mechanism to just provide the UID and to send out a notification using that framework written by US. So then your notification center would show what number or so. And then you will see whether you are mentioned and you will be linked to the file. And afterwards you want to make the clients a little bit smarter. So because the comments are plain text, I said the clients would need to be able to replace the UID with their display name. Also in mind, or having in mind editing the comment itself, next step would be the autocomplete feature. And this is that this would really become usable for the end user who kind of doesn't really know the UIDs. So the other things are just the needed fundamental base for this autocompletion. And yeah, so when this is done we could also, as a lock step, optimize a little bit to save some requests, for instance, for the clients to make it easier to work with at all. So the resources so far, there is an overview issue ticket where the steps are explained in detail. The first step actually is almost done a little bit and the smithing request will follow, hopefully, these days. And yes, the hack rooms, they are in the third floor. OK, thank you.